Bleeding and thrombotic events in adults supported with venovenous extracorporeal membrane oxygenation: an ELSO registry analysis

Abstract

Purpose: This study aimed at analyzing the prevalence, mortality association, and risk factors for bleeding and thrombosis events (BTEs) among adults supported with venovenous extracorporeal membrane oxygenation (VV-ECMO).

Methods: We queried the Extracorporeal Life Support Organization registry for adults supported with VV-ECMO from 2010 to 2017. Multivariable logistic regression modeling was used to assess the association between BTEs and in-hospital mortality and the predictors of BTEs.

Results: Among 7579 VV-ECMO patients meeting criteria, 40.2% experienced ≥ 1 BTE. Thrombotic events comprised 54.9% of all BTEs and were predominantly ECMO circuit thrombosis. BTE rates decreased significantly over the study period (p < 0.001). The inpatient mortality rate was 34.9%. Bleeding events (1.69 [1.49-1.93]) were more strongly associated with in-hospital mortality than thrombotic events (1.23 [1.08-1.41]) p < 0.01 for both. The BTEs most strongly associated with mortality were ischemic stroke (4.50 [2.55-7.97]) and medical bleeding, including intracranial (5.71 [4.02-8.09]), pulmonary (2.02 [1.54-2.67]), and gastrointestinal (1.54 [1.2-1.98]) hemorrhage, all p < 0.01. Risk factors for bleeding included acute kidney injury and pre-ECMO vasopressor support and for thrombosis were higher weight, multisite cannulation, pre-ECMO arrest, and higher PaCO₂ at ECMO initiation. Longer time on ECMO, younger age, higher pH, and earlier year of support were associated with bleeding and thrombosis.

Conclusions: Although decreasing over time, BTEs remain common during VV-ECMO and have a strong, cumulative association with in-hospital mortality. Thrombotic events are more frequent, but bleeding carries a higher risk of inpatient mortality. Differential risk factors for bleeding and thrombotic complications exist, raising the possibility of a tailored approach to VV-ECMO management.

Keywords: Bleeding; Stroke; Survival; Thrombosis; Venovenous extracorporeal membrane oxygenation.

Fig. 1

Flow diagram of patient selection for the primary analysis. Flow chart depicting inclusion and exclusion criteria. VV, venovenous; ECMO, extracorporeal life support

Fig. 2

Frequency of bleeding and thrombotic events during VV-ECMO. A Proportion of BTEs comprised of bleeding or thrombotic events. B Frequency of specific types of BTEs. BTEs, bleeding and thrombotic events; Oxy/pump failure, oxygenator/pump failure

Fig. 3

Association of bleeding and thrombotic events during VV-ECMO with in-hospital mortality. Association of BTEs with in-hospital mortality during VV-ECMO. AdjOR, adjusted odds ratio; CI, confidence interval

Fig. 4

Clinical and circuit characteristics associated with bleeding events. Factors associated with any bleeding events and medical bleeding (including intracranial hemorrhage, pulmonary and gastrointestinal bleeding). AdjOR, adjusted odds ratio; CI, confidence interval; ECMO, extracorporeal membrane oxygenation; PaCO₂, arterial partial pressure of carbon dioxide; P/F, PaO₂/FiO₂ ratio. The unit of observation for all models is the patient. *p < 0.05; **p < 0.01; ***p < 0.001

Fig. 5

Clinical and circuit characteristics associated with ischemic stroke and intracranial hemorrhage. Factors associated with ischemic stroke and intracranial hemorrhage. AdjOR, adjusted odds ratio; CI, confidence interval; ECMO, extracorporeal membrane oxygenation; PaCO₂, arterial partial pressure of carbon dioxide; PaO₂, arterial partial pressure of oxygen, P/F, PaO₂/FiO₂ ratio. The unit of observation for all models is the patient. *p < 0.05; **p < 0.01; ***p < 0.001.